摘要: 针对在线社会网络潜在社区难以检测的问题,根据在线社会网络的独有特性,提出一种基于用户紧密度的在线社会网络社区发现算法。创建初步用户图,依据节点属性相似性算法计算用户个体紧密度,基于共有邻居相似性算法计算用户社区紧密度,从而构造出完整用户图,利用层次聚类算法对完整用户图进行处理,发现潜在社区。实验结果表明,与NAS、CNS算法相比,该算法的社区凝聚度与正确率更高,分别达到0.67和97.1%。
关键词:
在线社会网络,
用户紧密度,
节点属性相似性,
共有邻居相似性,
社区发现,
层次聚类
Abstract: Aiming at the problem that it is difficult to detect the potential community of Online Social Networks(OSNs), based on the unique characteristics of OSNs, this paper proposes the new concept of user tightness, and designs a community detection algorithm based on it. It creates the initial user graph, computes user individual tightness based on node attribute similarity algorithm, and computes user community tightness based on common neighbor similarity algorithm, to create the integrated user graph, it processes the integrated user graph with hierarchical clustering algorithm, to detect the potential communities. Experimental result shows that compared with NAS algorithm and CNS algorithm, the detected communities of this algorithm have much higher degree of cohesion and accuracy, and reach 0.67 and 97.1%.
Key words:
Online Social Networks(OSNs),
user tightness,
node attribute similarity,
common neighbor similarity,
community detection,
hierarchical clustering
中图分类号:
熊正理, 姜文君, 王国军. 基于用户紧密度的在线社会网络社区发现算法[J]. 计算机工程, 2013, 39(8): 50-54.
XIONG Zheng-Li, JIANG Wen-Jun, WANG Guo-Jun. Community Detection Algorithm Based on User Tightness in Online Social Networks[J]. Computer Engineering, 2013, 39(8): 50-54.